I just spent an interesting afternoon trying to upgrade a 1.1 system to 1.3.
First, /var/lib/dpkg/available was corrupted because of some incorrect values in the Version - field (somehow they had gotten to the format of 1:1-2 or similar; bug report submitted). I fixed these by Then, there were lots of dselect problems, since I wanted to upgrade everything at once. ldconfig was still running in the old version (1.7.6), and libc wouldn't install because of this. In between, Perl got corrupted as well, so dpkg-ftp (which I used) didn't work either. Somehow, my libc *.deb file got truncated to 0 bytes. I ended up by copying some key files (libc5-*.deb, perl-*.deb) to the tmp directory via my non-Debian ssh (which still worked), then alternating between installing everything and configuring everything. After around three or four interations, this worked. Then, I rebooted to find that re-installing lilo had hosed my boot configuration completely. With this machine, I boot lilo from the master boot record of /dev/hda. My root partition resides on /dev/sdc2, and boot files on /dev/sdc1. After lots of fiddling around with the rescue disks, I finally managed to create a working boot disk and reboot into something remotely sane. After that, there was nothing that I hadn't experienced before (e.g. the X keyboard nationalization had suddenly stopped working - that bug report has long since been filed). If you have enemies, recommend that they should upgrade from Debian 1.1 to 1.3 this way. They will hate you forever for it. -- Thomas Koenig, [EMAIL PROTECTED], [EMAIL PROTECTED] The joy of engineering is to find a straight line on a double logarithmic diagram. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
